Who's afraid of knowing? Critical thinking, disclosure, and the reader in AI-mediated open scholarship

Engineering notes

Engineering notes will be added by the aipentium editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为大语言模型、生成式AI、ChatGPT相关技术、计算机视觉、深度学习等高价值论文补充中文说明。

Original abstract

A GCPA-SIDCER position paper in the Science, Policy, and Ethics in Conversation (SPEC) series. Drawing on the FSCI 2026 course 'Who's afraid of knowing?', which the author chaired, and on the reform of research assessment, the paper argues that artificial intelligence has become the epistemic environment of open scholarship rather than a tool used within it, and that its central risk is not falsehood but the premature closing of the space in which judgement, reflection, and conscience reside. It treats critical thinking, epistemic humility, and responsibility as core scholarly competencies rather than matters of compliance; it argues that AI disclosure should move from declaration towards provenance and verification in the reader's service, and should protect the person who discloses; it questions the haste with which scholarship has ruled that an AI cannot be an author, holding that what finally matters in a text is less whose name is attached than what happens in the mind of the reader; and it holds that research assessment must be reformed so that what is recorded and rewarded rewards human judgement rather than its removal. An afterword traces the four-year FSCI course series, 2023 to 2026, from which the paper grows. AI, provenance, and assurance statement. This paper grows from the FSCI 2026 course 'Who's afraid of knowing?', which the author chaired, and from the opening plenary, together with the three earlier courses in the series, developed for FSCI in 2023, 2024, and 2025, and the author's work in CoARA-ERIP, SE4RA, the EOSC-Future/RDA Artificial Intelligence and Data Visitation Working Group, the RDA Sharing Rewards and Credit Interest Group, and the International Data Policy Committee (IDPC) of the International Science Council's Committee on Data. The discussions among the teaching teams and participants across the four courses, the course notes, slides, and entry surveys, and handwritten and typed drafting, gave the paper its ideas and its structure. Large language models, including various models of Claude, ChatGPT, DeepSeek, and Microsoft 365 Copilot, contributed to the courses' development and discussion under the author's, faculties', and participants' directions. They also contributed to the development of this publication under the author's supervision, for structuring, drafting, and language editing. All content, claims, citations, and conclusions were set and verified by the named author, who takes full responsibility for the work. AI was not used to generate citations, data, or normative conclusions. The full structured statement is deposited with the record as a machine-readable object.

Links and sources

Comments